Data Mining Implementation On Java North Coast Weather Forecast Dataset Using C4.5 Algorithm

Abstract

Weather is one of the most influential things in everyday human life. Many activities carried out by humans cannot be separated from the prevailing weather conditions. Lately, there have been frequent irregularities in weather patterns that are not usual or can be said to be extreme. Therefore, observing the weather is very necessary to make predictions about the weather. The northern coastline is one of the most important routes in Java, especially the northern coast route in the Central Java, due to that information about weather forecasts on this route is needed. The purpose of this research was to obtain the most influence factors of weather changes. The data mining approach used in this research is decision tree method and C4.5 algorithm. From the test results of 2,400 weather forecast data taken from the accuweather site and divided into 2, namely training data as much as 1,680 data, the rest of testing data as much as 720 data, the results obtained from a decision tree with the root node is the humidity attribute with an accuracy rate of 81.94% which has been proven through rapid miner 9.10 tools.

Keywords: Weather Forecast, Data Mining, C4.5 Algorithm, Decision Tree